Discrimination Between Command and Drawing Strokes Based on Fuzzy Spline Curve Fragmentation and its Application to Sketch-Based CAD Interfaces

Sketch-based computer-aided design (CAD) interfaces based on fuzzy spline curve identifier allow users to create geometric drawings intuitively using penstrokes. However, they cannot create technical drawings because they lack command input. To naturally integrate command inputs into these interfaces, commands should be recognizable from penstrokes alone. Therefore, it is necessary to distinguish between command strokes (CS) for command input and drawing strokes (DS) for geometric-shape input in real-time without relying on contextual information or geometric features. We propose a CS/DS discriminator based on fuzzy spline curve fragmentation, introducing two techniques: fuzziness enhancement (FzEn) and phantom elimination (PhEl). We then illustrate the effects of FzEn and PhEl, experimentally showing that FzEn and PhEl significantly enhance the CS/DS discriminator performance. Furthermore, we apply the discriminator in sketch-based CAD interfaces and implement several trial command inputs. We demonstrate through drawing experiments that both the command and geometric-shape inputs are performed seamlessly to efficiently complete technical drawings.